Philosophy of Design: a meta-theoretical structure for design theory

Dr. Terence Love

PO Box 226 Quinns Rocks Western Australia 6030

Tel & Fax: +61 (0)8 9305 7629 Email: t.love@love.com.au

Pre-print copy

Love, T. (2000). Philosophy of Design: a Meta-theoretical Structure for Design Theory. Design Studies, 21(3), 293-313.

There may be differences between this pre-print and the final version.

For accuracy, please check the final version

Philosophy of Design: a meta-theoretical structure for design

theory

Abstract

This paper focuses on the structure and dynamic of theory in design research. Problems with existing theory are

explored, and a new meta-theoretical method is suggested for assisting the critical analysis, comparison and formulation

of design theories and concepts. This meta-theoretical method contributes to building a simplifying paradigm of design

research by providing a means to clarify the existing state of design theory in the field, to assist with the establishment

of coherence and compatibility between concepts in disparate theories, to validate theory and concepts, and to uncover

‘hidden’ aspects of design theories.

Keyw ords: philosophy of design, design theory, epistemology, ontology, design philosophy

Introduction

This paper is a contribution to the study of Philosophy of Design. In it a critical overview is taken of the

structure and dynamic of d esign theory. This overview leads to a method to assist with the critical analysis

of design theories and their associated concepts. If we consider ‘design theory’ as an artefact to be designed ,

this paper’s role is in the part of the design process often called ‘problem analysis’. Underpinning what is

presented in this paper is an assumption that the act of designing by human agents is central to the academic

study of design. This theoretical standpoint, that what is described as design, always implicates humans,

together with the understand ing that any theories, theorising or theory interpretation has meaning only in a

human context, is the basis for the analyses presented below.

Philosophy of Design is d ifferent from design philosophy. It is the d isciplinary equ ivalent of Philosophy of

Science, or Philosophy of Technology ------ whereas design philosophy is associated more with the

philosophical study of design method . Appropriate questions in Philosophy of Design are, for example;

‘What, in general, is design?’, or ‘What are the characteristics of a valid t heory of d esign?’, or ‘What are the

characteristic of a theory of objects being designed?’, or perhaps, ‘How might a theoretical design concept be

tested for coherency with other concep ts?’, or even ‘Should a theory of objects be part of design theory?’.

The term design philosophy has also been used to describe the study of such questions. The historically

intimate connections between design philosophy and the study of design methods 1,2 has, however, limited

the scope within which philosoph ical issues relating to design have been considered 3. Design Philosophy is

not, therefore, a suitable title for the philosophical stud y of all aspects of design theory because theories

related to design, and the concepts associated with them, cover a wider range of issues than the method s,

methodologies and techniques of design. The term, Philosophy of Design is used in this paper to describe this

wider philosophical inquiry.

In alignment with this philosophical perspective, the capitalised Design Theory is used here to refer to the

sub-d iscip line in which the role, valid ity, coherence and utility of theories and concepts pertaining to design

are researched . In this respect, Design Theory encompasses several other sub -d isciplines such as; Design

Science, Science of Design, Design History, Design Methods, Design Methodology. This is contrary to views

expressed elsewhere4-7 that design science (or the science of design) completely encompasses design theory and

the theory of technical objects, for example, in Hubka and Eder’s8 comprehensive development of a theory

of technical systems. From the above human-based position on design theory, the scientific viewpoint is but

one parad igmic perspective within which theorising about design may occur. Theorising about design goes

beyond the parad igms within which design is practised , and beyond those theoretical outlooks by which its

practice is researched , for example:

Coyne and his associates3,9-13 have written extensively on the app lication of a variety of post -

positivist outlooks to design research.

Sargent14 proposed a meta-theoretical argument that there cannot be a unifying d esign science

because there appears to be an incommensurability of viewpoints in design research

Broadbent15 d iscussed theory build ing in the study of design.

Thomas and Carroll 16 explored how design might be best conceived of in terms of an ind ividual’s

psychological perspective

Daley 17 analysed the role of objects in theories of creativity

Dilnot 18 investigated the limitations of definitions of design that excluded its social context

Liddament 19 d rew attention to the limitations that computationalist perspectives place on the

development of design theory, particularly in terms of ontology, epistemology and method ology

Galle 20 explored how the definition of ‘designing’ is influenced by the inclusion of human design

agents into its explanation (and helpfully clarified the language of ‘brief’, ‘representation’ and

‘artefact’)

Oxman21 created a new perspective on design education by focusing on the d ialectic nature of

designing, the associated theories of cognition and the epistemology of knowledge structures

All aspects of research into design involve conceptual abstraction and symbolic representation. In each of the

above examples, the theoretical abstractions of Design Science and other parad igms of design research are

subject to scrutiny, and , therefore must lie at a lower level of theoretical abstraction than the analyses of

them. In terms of sub-d isciplinary structure, this overarching role of Design Theory implies that it should be

viewed as being higher up the sub-disciplinary ‘tree’ of Design Research. Abstractions and representations

are fundamentally grounded in human values22-28, whether based on parad igmic assumptions25,26,29,

metaphors6,9,10, or reified conceptual frameworks30, and form part and parcel of theory-making about

design31. Taking a human-centred perspective on design research, therefore, also implies that Design Theory

subsumes most, if not all, sub-d isciplines associated with the study of design (theoretically, at least).

Confusion, conflation and multiplicity in Design Theory

The development of theories of design research has occurred in a piecemeal fashion, and four serious

criticisms may be made:

That there exists a substantial amount of confusion w ith respect to the underlying basis of many

theories, concepts and methods.

That in developing and validating theoretical aspects of the study of design, many writers are

unjustifiably conflating concepts d rawn from a range of sources.

That there exists an unnecessary multiplicity of design theories and concepts.

That the terminology of design research has become unnecessarily and unhelpfully confused and

imprecise by d int of the above points

In 1992, I attempted to collect together a glossary of the main theoretical terms of the design research

literature32, and found that it was an almost impossible task to do justice to the d ifferent variants of major

terms: there are almost as many d ifferent definitions of design and design process as there are writers about

design. Since the earliest attempts at formulating theories of design there have been many arguments

between proponents of d ifferent concepts and theories in which reconciliation happens by the proponents

realising that they were using the same word s or concepts d ifferently. Several researchers have pointed to

the problem that few of the definitions of terms associated with design research have been explicated

satisfactorily, for example:

Ullman33 claimed that ‘As research resu lts have been published , it has become obvious that the term

‘design’ has d ifferent meanings to d ifferent researchers. . . the field lacks a commonly accepted

description of design methods, types and theories’.

Pugh34 considered the stud y of the activities, philosophies, processes and products of design to be

confused .

Roozenburg35 pointed to the philosophical terminological and conceptual confusion surrounding the

abduction-deduction-induction aspects of innovative design thinking.

Eder 36 reported that a workshop aimed at developing a glossary of words and concepts relating to

the science of engineering design found problems where the same word s were used with d ifferent

concrete meanings, at d ifferent levels of abstraction, and where their meanings were dependent upon

cultural context.

Hubka and Eder8 noted the d ifficulty of clarifying definitions when creating novel theory.

Parnas and Clements37 argued that, in software design, precise definitions are often not provided , and

that ‘there are many terms used for the same concept and many similar but d istinct concepts

described by the same term’.

Talukdar, Rehg and Elfes38 claimed that neither practitioners nor researchers agree on what

constitutes design activity. They noted that Finger, Director of the Design Theory and Methodology

program of the US National Science Foundation, asked design researchers for their definitions of

design activity and received twelve d ifferent classes of answer.

French39 commented that there are as many ways of p roducing block d iagrams of design process as

there are of ‘tribal lays’.

In 1984, Cross2 identified four overlapping themes which also represent the chronological development of

research into design:

1. The management of the design process

2. The Structure of Design Problems.

3. The nature of Design Activities

4. "Reflection" on the fund amental concep ts of design

Two other themes that can be added to Cross’ list are:

5. Knowledge about the environment in which designing takes place

6. The knowled ge needed for designing, i.e. knowledge about objects and design processes (Hub ka and

Eder40)

Together, these themes can be viewed as d ifferent parad igms under which design researchers have

investigated design activities, and proposed theories and models to represent them. In 1993, Cross 1 reported

on theoretical developments over the ten years since his earlier review. This later review implies that a

sequential Kuhnian-style development of d ifferent sub-parad igms of design theory, in which the same terms

and concepts have been used d ifferently, has become the concurrent development of theory across all

parad igms. This would provide one explanation for both terminological and conflationary confusion in the

field .

Dixon7 confronted the confusion in design theory by regard ing all design research as being in a pre-theory

stage. His suggested , from a positivist scientistic position, that, as a matter of urgency, attempts are made to

establish scientifically testable theories of design. Dixon's outlook on the establishment of theories of de sign

may be subject to challenge by postpositivist academics (for example, Berger and Luckman 41, Coyne9,

Coyne and Snodgrass3, Coyne Snod grass & Martin12 and Margolis42), but his analysis of the potentially

impossible d ifficulties in establishing a general theory of design and his critique of attempts made so far,

were valid from either a positivist or post-positivist perspective.

It is d ifficult to prove that it is commonplace for theoretical aspects of the study of designing to be supported

by conflating concepts inappropriately. It is perhaps more fruitful to ind icate how conflation may lead to the

development of erroneous threads of theory or to faulty conclusions. The simplest case of conflation follows

the sequence:

‘A is related to B’ and ‘B is related to C’ therefore ‘A is related to C’

The valid ity of such an inference requires the correct identification and use of the characteristics of the

abstract entities (in this case A, B and C) and their relationships. A popular example of inappropriate

conflation is that of cat and dog equivalence:

‘A cat has four legs’ and ‘a dog has four legs’ therefore ‘a cat is a d og’

The construction of theory in the study of design sometimes brings forth similar problems with conflation,

for example,

‘Designers think’ and ‘cognitive psychologists study thinking’, therefore, ‘research into design lies w ithin

the d iscip line of Cognitive Psychology’

The issue of conflation is not simply a matter that the logic which has been used is faulty. It is that two or

more ideas have a similarity in some of their aspects which encourages their unjustified co -association and

improper conflation into some larger conceptual whole. This combined conceptualisation may then, in its

turn, be used as a faulty basis for the formulation of further theory, or to d raw incorrect conclusions from.

Whilst such obviously mud dled thinking is fortunately rare, the study of design and its associated theory is

conceptually challenging, and in theoretically d ifficult situations unjustified conflation can easily go

unnoticed by both author and critic alike.

Since the 1950s, the volume of design research literature has increased from a handful of books in the 60s to

the current level of several hundred books, articles and conference papers written each year. When Cross 1

identified the four sequential themes in design research mentioned earlier, he suggested that the

development of each was necessitated by the failure of its predecessor. The increase in the literature has

resulted in the number of theories and concepts in design. This high level of multiplicity of concepts and

theories in design research has several aspects:

Theories are speculatively proposed

Theories are generated from within a w ide variety of parad igms.

Theories and theoretical developments are not subjected to sufficient critical epistemological and

ontological attention.

There is little agreement on the most fund amental aspects of theory, for example, ‘What is meant by

design?’

Some useful theories that would help with the integration of Design Theory are ignored , partly

because they are ‘not invented here’, or perhaps because they lie outside what is seen as the province

of the stud y of design.

This problem of unbrid led abstraction, and its associated language needs, threatens to overwhelm design

research, and the protocols of academic research add to the problem. Whilst developing their own theories,

contemporary design researchers are sup porting their work by analysing the work of earlier theorists, each

of whom has been trying to provide theories of and about design, or critique the theories of those before

them. The next generation of theorists will in their turn be attempting to provide overviews of and theories

about current work. This continuous production of new abstractions by each new generation of design

researchers, and the requirement for terminology that d ifferentiates each new abstraction from earlier ones,

appears to be never end ing. The Ind ian literature of the Vedas provides a parallel to this situation in

describing how the earth is supported :

‘...on elephants, and they are supported on more elephants, and they on other elephants. Elephants on

elephants for ever....’

This is temporally-based conceptual development, in which new concepts and new terms are needed as time

goes by to describe patterns in the theories of previous generations. What is needed is some means of

structuring existing concepts and theories to bound the u nnecessary growth in abstractions and terminology

so that it is clearer to design researchers which concepts, theories and theoretical strands are pragmatically

more usefu l or better justified , and what their relationships are to each other. This is an important step in any

move toward the development of the ‘simplifying parad igm of design research’ that Cross 43 has identified

as important for the field .

The meta-theoretical method proposed in his paper is intended to hasten this process. The method provides

a bounded structure for the development of theories that reduces the need for new language and new

concepts to explicate the work of previous theorists. This simplification provides an underling structure that

enables the main focus of conceptualisation to be d irected towards theorising about the activity of designing,

the knowledge needed , the tools used , the information available, the management of the process, the

environment, and the specifications of products being developed .

Critical analysis and design theory

Critical analysis provides the basic tools for clarifying design theory because its purpose is the clarification of

relationships between ind ividual concepts and theories, and between these ind ividual concepts and theories

and their underlying assumptions. Critically analysing the concepts and theories associated with design is

not easy, however. In some cases, the researcher finds that an author has defined a concept to have a

particu lar meaning, and then used it later in th e same text with a subtly d ifferent meaning. In other cases,

authors confuse metaphoric meaning and literal meaning of terms. In other case still, a term may have been

applied simultaneously to d ifferent concepts, processes or activities. In many texts these problems are

compounded by confusion about the epistemological relationships between proposed ideas, and existing

concepts and theories. For example, it is clear that the study of the geometric properties of a particu lar

element of a design, such as a crankshaft, is d ifferent from trying to understand the cognitive mechanisms by

which a designer chooses such a shaft. Confusion arises in the literature, however, because the same words

and similar concepts are used in theoretically d ifferent circumstances. The following not untypical sentence

illustrates how ill-defined an apparently coherent phrase may be:

‘The design of the shaft depended on the decision making processes used by the d esigner.’

To be clear about the meaning of the sentence as a whole requ ires correct understand ing of the potential

meaning of each element in that sentence. This is not straightforward . For example, in the case of ‘decision

making processes’:

‘Decision making processes’ may refer to decisions based on basic models of shaft behaviour, for

example: models of force, stress or dynamics

‘Decision making processes’ may alternatively mean that the design is defined by the results of higher

order methods which use the output of, for example stress behaviour, to optimise shaft geometr y

Or, perhaps, it is the designer’s internal cognitive decision making processes that are being referred to

Perhaps ‘decision making processes’ refers to a process based on a wider internal perspective (or

world view) of the designer that includes their personal habits, values, beliefs about the world , the

effects of the designer’s environment, etc.

In the case of ‘the design of the shaft’:

‘The design of the shaft’ may mean the activity of designing the shaft

‘The design of the shaft’ may refer to an eng ineering d rawing of the shaft (or some other

communication or representation)

Perhaps ‘the design of the shaft’ refers to an actual shaft

By taking combinations of the possible meanings of each of the above two phrases in the above sentence it is

possible to see how easily confusion is brought into any d iscussion about designing, and how d ifficult it is,

at present, to clearly and unambiguously critically analyse even small amounts of design literature. The

meta-theoretical method proposed in this paper provides a means of resolving such conceptual tangles.

The semantic and conceptual confusion in design research is compounded by theoretical constructs at

d ifferent levels of abstraction having many similarities. There are several reasons that may be advanced to

explain this. From a Kuhnian perspective, researchers (often with little experience of designing products)

have a restricted range of conceptual tools, which when applied to research into d esign at any level of

abstraction produces similar sorts of an alyses. This is clearly true when a systems approach is applied . Both

designing and researching are human activities that are undertaken within particular cultural and

environmental ecologies which is likely to result in a uniformity of conceptualisation and language. This

argument is supported by existence, in the areas of History of Technology and Design History, of research

into the culturally based reasoning behind the development of particular artefacts or technologies (see, for

example, Margolin44).

Popper45 addressed the problem of theoretical confusion and valid ation with his analysis of the

relationships between the following three world s:

World 1 - Physical and material objects

World 2 - The subjective world containing mind s and their contents

World 3 - The objective world of theories, knowledge and problems

He argued that confusion arises when concepts from d ifferent worlds are conflated , and that it is better to

regard the analyses of each world as incommensurate. Phillips46 discussed Popper’s three world concept

with reference to research into educating native speakers of English to speak Russian. He asked how a

pedagogical theorist could valid ate a particular theory of language learning by testing a student. It is clear

that the linguistic skills and learning skills of a student could be tested by observing them answer, in Russian,

questions which are competently phrased in Russian. The internal subjective workings of the students mind ,

however, cannot be proven by these observations, and , more importantly, the information does not ind icate

the workings of the learning theory in question. If the researcher asked for subjective details from the student

it is not obvious how could the researcher prove them or use them to valid ate theoretical aspects of the

learning theory. For example, how could the researcher prove that the student was actually conversing in

Russian rather than using two acts of translation and thinking in English? Phillip’s examp le illustrates at

least one of the problems facing design theorists who are attempting to formulate well justified theory about

the internal creative processes of designers. (The reviewers of this paper have commented that this argument

is somewhat similar a position taken by Klaus47 on the relationships between ‘theory’, ‘method ’ and

‘object’.) Clearly, any theories about designing must comport well w ith our material and subjective worlds,

but the real test of theories lies in their valida tion and coherency with respect to other well supported

theories that are d rawn, not only from the field of design, but from the widest range of relevant theoretical

constructs across all d iscip lines. The proposed meta-theoretical method below provides a theoretical

structure for valid ating design theories and concepts in this manner.

Meta-theoretical analysis

Theories, concepts, assumptions and human values are stud ied and analysed as theoretical abstractions. They

are World 3 entities in Popper’s terminology. Although the practice of stud ying, generating, using and

criticising such abstractions is not commonplace in the contemporary d isciplines of science and technology,

such epistemological and ontological analysis is w idely used in other d isciplines and d ates back to the

earliest Greek philosophers. Whilst not at the forefront of academic consciousness, the structuring of

abstractions is a basic tool in most d isciplines. It can be seen perhaps most clearly in Mathematics - where

theorems depend on axioms (both abstractions), and in Philosophy - especially in the field of Logic, which is

concerned with the manipu lation of abstract entities and the verification of logical relationships between

them.

Discip lines that involve practical human action, particularly the Social Sciences such as, Anthropology,

Ethnography and Sociology, have a significant focus on the epistemological and ontological analysis of their

theoretical abstractions because their theoretical found ations are not amenable to Cartesian validat ion. This

lack of Cartesian validation means that the development of empirical research programs requ ires a critical

identification of the assumptions and implications of abstractions reaching back to an understand ing of the

ontological and epistemological bases of such research . The development of Grounded Theory by Glaser and

Strauss48 and its method s of validation are a typical case. A human -based perspective on design research

implies that it should be viewed in the same light as these other d isciplines that involve research into human

action and , in consequence, needs a similar level of attention to ep istemological and ontological issues

relating to theories involving humans in design.

There are many d ifferent means by which similar attention to epistemological and ontological clarity could

be brought to design research, and some are more appropriate than others. In this paper, a meta -theoretical

approach is used because it is just post-positivist and aligns well with Science and with a constructivist

approach to human knowing. The approach is meta-theoretical because the analysis of a particu lar set of

theories can only be done at a higher level of abstraction, and because concepts and theories exist in meta -

level relationships to each other. The basis of meta-theoretical analysis is that it provides a means to analyse,

relate, position and valid ate concepts and theories that are in and from d ifferent theories and theoretical

stances.

Metaphors of design as meta-level abstractions

It is not just the ‘hard’ theories of engineering design that are abstractions in the sense used above. It applies

to metaphors and other mental constructs also. Metaphors are used describe phenomena or patterns of

mental constructs as if they were something else------something more familiar------ and may be used consciously

or unconsciously. This is relevant to design research because attempts to describe the process of design have

been generally metaphoric in nature, and this is especially true of general theories of d esign. These

metaphors are usefu l in that they enable the grouping of concepts at lower levels of abstraction into

memorable patterns. Such groupings or meta-abstractions allow us to place concepts relative to each other,

describe relationships between them and fit new concepts into an overall theory. Each metaphor, however, is

both limited in its scope and limiting for its users. Whilst metaphors provide a basis for checking the internal

consistency of a grouping of concepts they also exclud e other ways of describing and analysing phenomena

and are limited by them not being literal descriptions. This in turn may cause d ifficulties in includ ing new

insights and information. Coyne and Snodgrass3 used this metaphoric view of design to d iscuss the

limitations of the overarching metaphors of design process in how design problems are formed and

addressed . A lack of clarity about abstractions, meta-abstractions and metaphors along with an unconscious

use of metaphors or meta-abstractions of design is likely to result in semantic d ifficulties and theoretical

confusion.

The use of meta-theoretical analysis

Abstractions are closely tied to human assumptions (which in their turn, are also abstractions) about how the

world is. Every theory or concept is dependent upon a variety of other abstractions at d ifferent levels of

abstraction than itself, and for its valid ity it must be seen to be coherent w ith them. Abstractions relating to

any situation lie in a meta-theoretical spectrum in which the lowest level of this spectrum refers to the

sensual ‘concrete’ world as experienced through our senses, and the most abstract higher levels refer to the

ontological world of assumptions about reality, personal values and value systems. In between lie all other

abstractions, each grounded in human experience and conceived on the basis of our ontological assumptions

about existence.

To give an example of this hierarchy. Consider the concept of stress as an abstraction in engineering

mechanics. Contrary to the beliefs of many engineers, stress does not exist as a physically perceivable

phenomena (except, perhaps, mental or psychological stress!). It is a theoretical abstraction depending for its

meaning on the lower level abstractions of force and area. Usually stress is defined as the force transmitted

per unit area. The abstractions force and area in their turn depend for their meanings on our more concrete

conceptualisation of our observations of the movements of physical objects.

The metaphor or parad igm within which we can conceive mechanical stress also, however, depends upon

the human values espoused in higher level abstractions about existence such as:

Assumptions about the consistency of the universe

Assumptions that the world can be mathematically modelled

Reliabilistic beliefs about the constancy of the world and our perceptions of it (see for example Levin 49

)

A religious belief system which allows us to make such models without fear of d ivine retribution.

The abstraction ‘stress’, therefore, lies in a hierarchy of abstractions. It depends upon both lower and higher

level abstractions. It not only depends upon these other abstractions and its coherency with them for its

meaning but it is validated (or not) by them.

To summarise: Theories and concepts are abstractions (Popper’s World 3 entities). These abstractions lie

within a co-dependent hierarchy. At the lowest level of abstraction is the d irect perception of reality, where

information is available to each ind ividual through their perceptive senses. The highest level of abstraction is

concerned with the beliefs and values associated with fundamental issues of existence. Between d irect

perceptions of 'reality' and beliefs about 'what is fundamental about existence' we have the everyday

abstractions which are the stock in trade of communication, reflection and theorising in such d iverse

occupations as journalists, artists, scientists, technologists and academics.

A meta-theoretical structure for Design Theory

One simple solu tion to problems of confusion and conflation of the concepts, theories and terminology in

design research is to take a meta-theoretical perspective, use a critical framework for analysis, and create a

structure that enables elements of d ifferent theories and concepts to be located relative to each other. The

most obviously usefu l method is to use a meta-theoretical structure based on levels of abstraction because it

offers a means of classification that is hierarchical and relatively independent of the domain -based meanings

associated with each theoretical element. This method provides a straightforward means of clarifying and

externalising many of the hidden dependencies between abstractions in Design Theory.

The meta-theoretical taxonomy proposed below also offers an opportunity for using rationalist methods to

inspect and critique the theoretical frameworks that research is undertaken within, and to explore the

bound s that particular ind ividual and cu lturally formed realities p lace upon theory -making. In this meta-

theoretical taxonomy, the levels of abstraction or meta-theory are based on a hierarchical form. The lowest

level of abstraction refers to an ind ividual’s d irect, sensual interaction with the world . The highest level is

concerned with human values, assumptions about existence and the implications of those assumptions. Thus

the taxonomy ranges from ‘our perception of reality’ to our questioning ‘What is reality?’.

It is emphasised that although this taxonomy was developed as a basis for categorising d ifferent theoreti cal

abstractions about designing, the focus is not on the content of the abstractions or theories. This is a

taxonomy of abstractions and theories in terms of their theoretical behaviour that is a stud y of theory qua

theory. The relationships which are important in this taxonomy are those of theoretical and conceptual

definition, and necessary assumption. To give an example, the meta -theoretical level of ‘Theories relating to

mechanisms of choice’ is concerned with choices about particular theoretical elem ents that each have

particu lar behaviours. At root, the description of these elements and their behaviours depends on empirical

experience (level 1). These theories about choice also d epend (however unconsciously to users and theorists)

on various privileged assumptions at higher orders of abstraction, such as assumptions, and presumptions

about design method and process, or even more abstractly, about what designing is, or more abstractly still,

about what the world is.

Meta-theoretical structure for classifying abstractions of design theory

Each level in following hierarchy contains: the number and name of the abstraction level, a short description

and illustrations or examples.

1. Direct percept ion of realit ies - This is the level at which we ‘sit on chairs’, ‘watch sunsets’, ‘hear the

sound of a bird ’ ... - ‘The woodworker feels the movement of the hammer as the nail is driven.’

2. Descript ion of Objects - The level that encompasses simple descriptions of objects, processes and

systems. - ‘a vacuum cleaner’, ‘a car body’, ‘a groyne’, ‘a typeface’, ‘a database’… ‘The woodworker uses a ‘claw

hammer’ rather than a ‘chisel’.’

3. Behaviour of Elements - The level at which the behaviour of elements which may be incorporated into

objects, processes and systems is described . For example, ‘a camshaft rotates at 600 rads/sec’, ‘headline type

needs to be set closer than body text’, ‘the lower windows need to offset the visual weight of the portico’, ‘the

melody returns to the tonic’. ‘The hammer is made up of two parts; a head and a handle… ‘The correct angle

between the handle and the face of the hammer head is necessary for nails to be hammered in straight.’

4. Mechanisms of Choice - The level of descriptions about the way that choices are made between

d ifferent objects, processes, or systems, and how solu tions are evaluated . For example, ‘Why does a

woodworker choose a claw hammer rather than a sledge hammer for hammering a small nail?’

5. Design Methods - The level in which theories about and proposals for design methods an d techniques

are described . - The theories about designing wood artefacts. ‘How does one design a chair?’

6. Design Process St ructure - The level that includes the theories about the underlying structure of

design process, and the influences of domain, culture, artefact type and other similar attributes and

circumstances. For example, ‘What are the processes underlying the design of Polynesian catamarans?’

7. Theories about the Internal Processes of Designers and Collaborat ion - This level includes the

descriptions of theories about the reasoning and cognition of ind ivid ual designers, of negotiated

design in collaborative design teams, and of cu ltural d esign effects on designers’ output. For example,

‘How did Mackintosh design furniture?’ ‘What communication is necessary between the different designers of

timber framed housing?’

8. General Design Theories - This is the level that is concerned with the details of those general theories

which seek to describe the whole activity of designing and its relationship to th e objects involved . For

example, ‘The activity of designing a boat, or a turbine, or a comic strip can be described as follows....’

9. Epistemology of Design Theory and the Theories of Objects - This is the level that contains those

analyses and d iscussions about the critical stud y of the nature, ground s, limits and criteria or valid ity

of design knowledge. - ‘What is a theory of design?’, ‘What does it include and exclude?’, and ‘On what

assumptions is this theory based?’

10. Ontology of Design - The philosophical study of the ontological basis for design theory and the

activity of designing. It is at this level where human values, and the values and fundamental

assumptions of researchers, are included in critiques of theory. For example, ‘Which human values and

assumptions effect the design of new legislation for narcotics?’, ‘Are the methods of evaluation used to choose

between different design alternatives consistent with the ethical proscriptions of the relevant professional

bodies?’, ‘What is reality?’, ‘What is existence?’.

The above taxonomy has been structured hierarchically in a manner which separates d ifferent groups of

abstractions, as found in the literature on design research, into a dependent sequence. Abstractions in levels

2 to 10 provide theories and patterns of concep ts at lower levels. Abstractions in levels 1 to 9 of the taxonomy

are based on assumptions d rawn from following levels. Abstractions at higher levels provides assumptions

for preced ing levels. Clearly there is much interaction between research and theory generation done at each

level of abstraction and other levels. Where new concepts are proposed at any level, appropriate language is

needed to separate its concepts from that of other levels of abstraction.

At first glance,

Levels 2 and 3 relate to objects

Levels 4---7 relate to design process

Levels 8---10 relate to philosophical matters

On this basis it may be argued that the theory categories relating to objects (levels 2 and 3) and the theory

categories relating to design process (levels 4---7) form two parallel but co-ord inated streams. This may be a

useful identification of the foci of research effort, but in terms of theoretical relationships between

abstractions the matter is d ifferent. Theories relating to design process app ly to other less abstract elements or

entities and their behaviour. It is levels 2 and 3 that contain the descrip tions and details of behaviour of these

entities, and hence, in terms of abstraction, there is a hierarchical dependence between levels 4-7 and 2-3. It is

irrelevant that some of these level 2 and 3 entities or elements, particu larly the non -physical ones, are

difficult to conceive, or in human terms ‘very abstract’ because it is matters of theory structure and dynamic

which are being addressed rather than cognitive d ifficulty.

The elementary verbal or literal ‘descriptive tokens’ of level 2 do not necessarily refer only to the physical.

Complex theoretical abstractions may also be tokenised similarly; for example, d ifferent logic chip

architectures may be given descriptive tokens and dealt with as abstractions in a similar manner to more

physical phenomena. This tokenisation also applies to elements of design theories. This is as it should be,

because meta-theoretical analysis is app licable to any abstraction or abstraction structure where the focus of

the study is on theory-making qua theory-making.

In the above meta-theoretical structure, d ifferent sub-d isciplines or fields of design research have a d ifferent

balance of activity at each level. For example, the main theoretical abstractions relating to Engineering and

Graphic Art, might well occupy predominantly d ifferent niches, or have d ifferent d istributions in the above

hierarchy. Both will be d ifferent to the d istribu tion of the more abstract aspects of Design Science (or Science

of Design) which would be found mainly in levels 9 and 10 (epistemology and ontology).

Phillips46 evaluated several contributory postulations on the development of knowledge and noted the

weight of academic opinion that all theory is unprovable in isolation but depends for its existence on a wider

structure or theoretical ecology (see also, Stegmüller 31 and Murray50). It should be expected that successful

general theories of design would provide a complete range of coherent concepts at all levels. In add ition,

each ind ividual concept or theoretical element should also, potentially at least, be a part of a general theory.

For each theoretical element or theory to be coherent, and to stand up to the rigours of comparative and

critical analysis, it must be a part of a web of theory that ranges from d irect perception of reality to

assumptions about existence itself. To repeat, theoretical elements and concepts should form coherent chains

that are represented in all levels of the meta-theoretical hierarchy.

The use of the meta-theoretical structure

The above structure provid es the means to improve conceptual and linguistic clarity by decomposing

d iffering design theories and concepts in to their relevant contributions at each of the d ifferent levels of

abstraction. This decomposition provides a powerful means of comparing and contrasting d ifferent theories,

concepts and metaphors of design theory by exploring the d ifferences and similarities between related

elements at each meta-theoretical level. The following simplified example shows how two metaphors of

design (Design as Information Processing and Design as a Creative Process) can be decomposed into their

contributions at each level of the meta-theoretical hierarchy to provide the conceptual detail that enables

their critical analysis.

Design as Informat ion Processing

Information processing is the most common theoretical perspective on design found in the contemporary

literature of design research. This perspective is closely approximated by that of Suh 5.

1. Direct percept ion of realit ies - Receiving and sending information. The effect on the user. Feelings of

‘lack of information’ or ‘being overwhelmed by too much infor mation’ or ‘wondering where to find

appropriate information’.

2. Descript ion of Objects - Descriptions of information sources such as; books, computer files, expert

knowledge. Descriptions of information types such as; ‘fluid viscosity’, ‘Pantone codes’, ‘customer

needs’ and ‘product specifications’. Descriptions of information flow processes such as; ‘telephoning’,

‘DXF file transfer’, ‘interaction with stakeholders’, ‘human -machine interactions’, ‘design team

collaboration’.

3. Behaviour of Elements - The behaviour of elements of the designed artefacts are d escribed in terms

which allow for convenient information processing. For example, the use of mathematically based

models rather than scale models.

4. Mechanisms of Choice - The relevance of decision making methods and evaluatory techniques is

dependent upon their communication efficiency, data integrity and data valid ation. The relevant

fields of study which relate to the reasoning behind design choice would include; communication

theory, artificial intelligence, cognitive psychology of decision-making, systems analysis and

information processing.

5. Design Methods - Design is seen as the cod ification, selection and management of information. The

characteristic design method is the use of information selecting algorithms.

6. Design Process St ructure - The design of an artefact is seen as an information selection and

management process. The underlying metaphor has the following aspects:

The world can be cod ified in a form suitable for information processing

This coded information can be categorised

The correct information can be supplied to the designer

The designer can use an information processing algorithm to ‘design’

A description of the designed artefact can be provided for manufacture

The manufacture of the artefact can be achieved from the supplied information without ad -hoc

interpretation.

7. Theories about the Internal Processes of Designers and Collaborat ion - The designer is seen as a

machine capable of rationally selecting and connecting together elemental information to satisfy a set

of constraints. It is assumed that human expertise can be mathematically modelled . Interaction

between design team members is seen as the communication of information.

8. General Design Theories - ‘Design as information processing’ lies within an atomistic and

deterministic framework. It is presumed that design can be automated and that the human designer is

only necessary to oversee that process.

9. Epistemology of Design Theory - The test of whether particu lar design information, methods and

theories are satisfactory is whether they transform and transfer information in the manner in which

they were conceived .

10. Ontology of Design - ‘Design as information processing’ assumes that information is ‘value neutral’.

The universe is presumed to be consistent. It is assumed that it is possible to model reality exactly.

‘Reality’ is taken to be cod ifiable in such a way that information has the same meaning for any person

using it. Cultural factors and human values are quantified .

Design as a Creat ive Process

Creativity that emphasises the intuitive internal creative processes of the designer forms the basis of this

alternative metaphor of design research. This perspective has been unfashionable in engineering design

research for some time. Amabile51 noted that research into creativity has been mainly undertaken within

Psychology, and even there it is relatively unrepresented . This creative perspective on the design process is

exemplified by Glegg52.

1. Direct percept ion of realit ies - Observation through the senses------mainly sight in most d iscip lines. The

representation of observations about bodily sensations and kineasthetics as ‘feelings’.

2. Descript ion of Objects - Objects are typically described ad jectivally rather than by using simple noun

descriptions. Some examples are: a ‘strong’ red , a ‘pretentious’ vestibule, a ‘convoluted’ melody, an

‘unnecessarily complex’ mechanism, an ‘elegant’ solution.

3. Behaviour of Elements - The main focus is on the interrelationships between elements. Although

elements have intrinsic characteristics their properties are more commonly defined by other elements

and external influences. To give some examples: from engineering design - ‘a gearbox may be well

matched to an engine’, from typographic design - ‘this subhead type is more dominant than the main

headings’, from graphic design - ‘the vertical line divides the picture unequally’, from architecture - ‘the

low roofline helps the bu ild ing to blend into the land scape’, from pedagog ical design - ‘group exercises

enable students to participate more fully in their education.

4. Mechanisms of Choice - Where design is seen as a creative process the dominant mechanism of

decision-making and evaluation is the use of ‘feeling’. Stud ies of the ‘correctness’ of design decision -

making based on feeling are usually historical and cultural. Such analysis depend s also on an

assessment of feeling and is usually justified by casuistic means. The type of questions which might be

asked in this metaphor are; ‘Which particular outlook should be used for designing in this case? or

‘How can Roman form and Jazz imagery be juxtaposed satisfactorily’, or ‘How can a successful

compromise be made in placing the camshaft?’. The answers to such questions depend upon human

values and as such lie in the domain of ‘feeling’.

5. Design Methods - A range of methods have been developed to facilitate the designer’s use of the right

hemisphere of the brain. Such methods include associative and analogical techniques such as

Synectics, Mind maps and Brainstorming. Many of these methods are also intended to d iscourage

analytical thinking or use of the left hemisphere of the brain. Other methods provide guidance in

specific domains for visual creativity and manipulation of concep ts. These techniques include

concepts of visual balance, the flow of form, repetitive elements and geometrical transformations. All

creative design methods necessarily depend on a sufficient base of experience resid ing within the

designer(s).

6. Design Process St ructure - Design process models are often similar to more technically based process

models. Many models use a systems based description of the form analysis - synthesis - evaluation with

evaluation feed ing information back into the other elements. Models based on a Romantic metaphor of

design often omit the evaluation feedback loop, presuming that the ind ividual genius of the designer

is the measure of the quality of the design. The ‘creative’ aspect of the design process is seen as

‘intu itive’ or mysterious and is the most dominant aspect of the process; with all other process

elements having a supporting role.

7. Theories about the Internal Processes of Designers and Collaborat ion - Some theories are Romantic in

style emphasising the creative genius of the ind ividual. Whilst attempting to d iscuss creativity

phenomenologically the underlying assumption is of creativity as a mysterious process. Other theories

try to explain creativity as a function of particular biological and psychological processes. Descr iptions

of ind ivid ual designer’s creative processes refer to the designer’s intuition, experience, feelings and

style together with the domain’s trad itions. Collaboration between designers is seen as a process of

trying to communicate nuances of feeling.

8. General Design Theories - Design is seen as a creative activity. Other aspects of d esign process are

subord inate to this.

9. Epistemology of Design Theory - Assessment of the valid ity or coherency of design information,

methods and theories is seen to be par t of the intrinsic creative activity of the designer or design

theorist. Domain based critics attempt to provide external comment on the completed works of

designers.

10. Ontology of Design - There are many ontological bases espoused by those who view design as a

creative process. This metaphor of design includes human values, attitudes and assumptions. It d oes

not depend on any particular world view except in the assumption that design is not deterministic, i.e.

that it is not possible to red uce creativity to a set of algorithmic steps. It further assumes that design is

an exclusively human activity which cannot be automated .

Using the results of the meta-theoretical decomposition

The above method of meta-theoretical analysis shows many d ifferences between the two metaphors ------ in

spite of the limited and sketchy nature of the deconstructions. The method also ind icates the theoretical or

conceptual range of each metaphor and helps avoid the problem caused by proponents of a metaphor

extending it by broadening its meaning without explaining the theoretical consequences. For example,

Information Processing might be extended by claiming that ‘everything can be expressed as information’, or

Creativity might be extended by insisting that every perception, action or thought is ‘new’ and hence a new

creation. This technique of extending a metaphor may be useful temporarily for generating new insights, bu t

it commonly leads to theoretical and terminological confusion in the literature. In both cases, these changes

would grant greater powers to each metaphor, but if this increase in the range of a metaphor was given

without the necessary analysis as to whether it was justified then it would simply be an example of

theoretical ‘d ouble-speak’. The meta-theoretical method helps avoid this and similar problems because it

shows the actual theoretical structure that is ascribed to a metaphor regard less of any transitional semantic

reinterpretation of its title. Meta-theoretical analysis publicly lays bare the agreed underlying facets of any

design theory so that when a metaphor, theory, concep t, or other abstraction is proposed , or changed , it can

be critiqued not only in terms of its own content but also in terms of its conceptual placement and its

relationships. That is, the meta-theoretical structure provides the theoretical framework within which it is

possible to not only ask, ‘What d o you mean by this id ea?’, but also to ask, ‘What other concepts and theories

is it related to and how?’, ‘What assumptions does it depend on?’, ‘What implications has it for the meanings

of other concepts or theories?’ and ‘How coherent is it with other concepts and abstractions?’.

Summary

This paper lies within the field of Philosophy of Design rather than Design Philosophy. In the paper, the

problems in Design Theory that are caused by widespread confusion and conflation and the unnecessary

multiplicity of theoretical concepts are addressed via a meta -theoretical analysis, Popper’s three world view

and metaphor. A meta-theoretical method that assists with the move toward a simplifying parad igm of

design research is proposed and an example of the use of this meta -theoretical method is given in which the

underlying theoretical basis of Design as Information Processing is compared with that of Design as a Creative

Process.

Conclusion

Confusion, semantic proliferation and lack of coherency in design theory may be reduced by the app lication

of meta-theoretical analysis. The meta-theoretical method described above assists the critical and rad ical

assessment of theories, concepts and other abstractions in Design Research. The method is simple and

straightforward to use, and provides a basis for limiting the generation of unnecessary terminology in

design research. The meta-theoretical method also provides a structured means of identifying which

elements of design theory might contribu te to a truly ‘simplifying parad igm’ of design research, and which

potentially simplifying parad igms are epistemologically inappropriate.

Acknowledgements

I am indebted to Dr. Jane Howarth of Lancaster University and to the late John Woollatt for d iscussing the

philosophical aspects of design research with me. I thank the reviewers of this paper for their constructive

advice and guid ance. My thanks also go to the Engineering department of the University of Western

Australia, and Drs. Laurence Spencer and Malcolm Hollick in particular, for their support.

References

1 Cross, N 'Science and Design Methodology' Research in Engineering Design Vol 5 (1993) pp 63-69

2 Cross, N (ed) Developments in Design Methodology John Wiley, UK, (1984)

3 Coyne, R and Snodgrass, A B Problem Setting within Prevalent Metaphors of Design - Working Report

Faculty of Architecture, University of Sydney, Sydney, (1992)

4 Simon, H A The Sciences of the Artificial 2nd edn, MIT Press, Cambridge, Ma, (1984)

5 Suh, N P The Principles of Design Oxford University Press, New York, (1990)

6 Altschuller, G S Creativity as an exact science Gordon and Breach Science Publishers, London, (1984)

7 Dixon, J R On a research methodology towards a scientific theory of design. In S L Newsome, W R Spillers

and S Finger (eds) Design Theory '88, Springer-Verlag, Berlin, (1988)

8 Hubka, V and Eder, W E Theory of Technical Systems Springer-Verlag, Berlin, (1988)

9 Coyne, R D 'Objectivity and the design process' Environment and Planning B: Planning and Design Vol 19

(1990) pp 361-371

10 Coyne, R D and Yokozawa, M 'Computer Assistance in Designing from Precedent' Environment and Planning

B: Planning and Design Vol 19 (1992) pp 143-171

11 Newton, S and Coyne, R D 'Impact of Connectionist Systems on Design' Knowledge Based Systems Vol 5 No

1 (March) (1992) pp 66-81

12 Coyne, R D, Snodgrass, A and Martin, D Metaphors in the design studio - Working Report Faculty of

Architecture, University of Sydney, Sydney, (1992)

13 Coyne, R D and Snodgrass, A 'Is designing mysterious? Challenging the dual knowledge thesis' Design

Studies Vol 12 No 3 (1991) pp 124-131

14 Sargent, P 'Design Science or nonscience' Design Studies Vol 15 No 4 (1994) pp 389-402

15 Broadbent, G Design and Theory Building. In N Cross (ed) Developments in Design Methodology, John

Wiley, UK, (1984)

16 Thomas, J C and Carroll, J M 'The Psychological Study of Design' Design Studies Vol 1 No 1 (1979) pp 5-11

17 Daley, J 'Design Creativity and the Understanding of Objects' Design Studies Vol 3 No 3 (1982) pp 133-137

18 Dilnot, C 'Design as a socially significant activity: an introduction' Design Studies Vol 3 No 3 (1982) p 139–146

19 Liddament, T 'The computationalist paradigm in design research' Design Studies Vol 20 No 1 (1999) pp 41-56

20 Galle, P 'Design as intentional action: a conceptual analysis' Design Studies Vol 20 No 1 (1999) pp 57-82

21 Oxman, R 'Educating the designerly thinker' Design Studies Vol 20 No 2 (1999) pp 105-122

22 Dewey, J How we think: a restatement of the relation of reflective thinking to the educative process D.C. Heath

and Company, New York, (1933)

23 Dewey, J Art and Experience Capricorn Books, (1959)

24 Feyerabend, P Against Method New Left Books, London, (1975)

25 Kuhn, T S The Structure of Scientific Revolutions Chicago Press, Chicago, (1962)

26 Guba, E G The Alternative Paradigm Dialog. In E G Guba (ed) The Paradigm Dialog, Sage Publications,

London, (1990), pp 17-27

27 Lincoln, Y The Making of a Constructivist. In E Guba (ed) The Paradigm Dialog, Sage Publications Inc.,

California, (1990)

28 Kitchener, K S and Brenner, H G Wisdom and Reflective Judgement: knowing in the face of uncertainty. In J

R Sternberg (ed) Wisdom: its Nature, Origins and Development, Cambridge University Press, Cambridge,

(1990), p 212–229

29 Smith, J K Alternative research paradigms and the problems of criteria. In E G Guba (ed) The Paradigm

Dialog, Sage Publications, London, (1990), pp 167-187

30 Nideau, R L Mind, Machines and Human Consciousness Contemporary Books, Chicago, (1991)

31 Stegmüller, W The Structure and Dynamics of Theories Springer-Verlag, New York, (1976)

32 Love, T Social, Environmental and Ethical Factors in Engineering Design Theory: a Post positivist Approach

Praxis Education, Perth, Western Australia, (1998)

33 Ullman, D G 'A Taxonomy for Mechanical Design' Research in Engineering Design Vol 3 (1992) pp 179-189

34 Pugh, S 'Engineering Design - Unscrambling the Research Issues' Research in Engineering Design Vol 1 No 1

(1990) pp 65-72

35 Roozenburg, N On the Logic of Innovative Design. In N Cross, K Dorst and N Roozenburg (eds) Research

into Design thinking, Delft University Press, The Netherlands, (1992)

36 Eder, W E Report on workshop W3. In V Hubka and W E Eder (eds) Schriftenreihe WDK 7 Results of ICED

81 (Rome), Heurista, Zurich, (1981)

37 Parnas, D L and Clements, P C 'A Rational Design Process: How and Why to Fake it' IEEE Transactions on

Software Engineering Vol SE-12 No 2 (1986) pp 251-257

38 Talukdar, S, Rehg, J and Elfes, A Descriptive Models for Design Projects. In J S Gero (ed) Artificial

Intelligence in Engineering Design, Computational Mechanics Publications, Avon, UK, (1988)

39 French, M J Conceptual Design for Engineers 2nd edn, Design Council, London, (1985)

40 Hubka, V and Eder, W 'Design Knowledge: Theory in Support of Practice' Journal of Engineering Design Vol 1

No 1 (1990) pp 97-108

41 Berger, P and Luckman, T The Social Construction of Reality Penguin Books, England, (1987)

42 Margolis, J The Technological Self. In E F Byrne and J C Pitt (eds) Technological Transformation: Contextual

and Conceptual Implications, Kluwer Academic Publishers, Netherlands, (1989), pp 1-16

43 Cross, N Research in design thinking. In N Cross, K Dorst and N Roozenburg (eds) Research in Design

Thinking, Delft University Press, Delft, (1992), pp 3-10

44 Margolin, V 'Design History or Design Studies: Subject matter and methods' Design Studies Vol 13 No 2 (1992)

pp 104-116

45 Popper, K Unended Quest Open Court, Illinois, (1976)

46 Phillips, D C Philosophy, Science and Social Inquiry Pergamon Press, Oxford, UK, (1989)

47 Klaus, G Kybernetik in Philosophischer Sicht 4th edn, Deitz-Verlag, Berlin, (1965)

48 Glaser, B G and Strauss, A L The Discovery of Grounded Theory: Strategies for Qualitative Research Aldine,

Chicago, (1973)

49 Levin, M 'Reliabilism and Induction' Synthese Vol 97 No 3 (1993) pp 297-334

50 Murray, E L Imaginative Thinking and Human Existence Duquesne University Press, Pittsburgh, (1986)

51 Amabile, T M The Social Psychology of Creativity Springer-Verlag, New York, (1983)

52 Glegg, G L The Design of Design The University Press, Cambridge, UK, (1971)